Make Apps with ChatGPT and Generative AI

1. App Development Introduction

This brief introductory video outlines the contents of the course, the concepts to be covered, and the features, pros, and cons of ChatGPT. It also outlines briefly the concepts covered in the course.

2. Roadmap: Building ChatGPT Apps

This video explains the roadmap to the course and the upcoming concepts and what is covered. The video explains the approach of the course, what you will learn in a step-wise manner, and other course content.

3. Examples: ChatGPT Solutions

This lesson demonstrates a few examples of successfully created ChatGPT applications and solutions, including the GitHub Copilot, Keeper Tax, Viable, and Duolingo.

4. Tech Stack: ChatGPT Apps

In this video, we will understand how ChatGPT works and creates applications or solutions theoretically, including the frontend, the backend, a wrapper API, the OpenAI API, and the GPT-3 Model.

5. Instructor Overview

This video is an introduction to the instructor, Henry Habib, and outlines his experience, qualifications, and his journey in the online training portals.

6. Keys to Success: ChatGPT App Development

This video by the instructor provides tips and tricks on how to complete the course successfully and make the best use of the learning process and the application development concepts.

1. ChatGPT for Problem Solving

In this video, we will understand how ChatGPT helps provide problem-solving solutions and explains how the concept works. We will understand the basics, how it fits into OpenAI, what the different models are, pros and cons, and how to integrate ChatGPT into our lives.

2. OpenAI

This video discusses about the San Franciscan AI deployment company, OpenAI. We will understand the mission of the company, what it delivers, and what cutting edge technology it is coming up with. We will understand the different models available from OpenAI.

3. GPT3 Models: Pros and Cons

In this lesson, we will explore the different GPT-3 models available, understand their features and capabilities, and also the advantages and disadvantages of the Davinci, Curie, Babbage, and Ada models. How to select the model of choice based on your requirements is discussed.

4. GPT3 Davinci: Capabilities

This video demonstrates the Davinci model and explains to us the capabilities of the Davinci model and what things it can do. You will learn about its classification and categorization abilities, content generation, conversation, and prompt-responses.

5. Optimizing Prompts

In this video, we will understand the other key concepts of ChatGPT before we begin creating models. You will learn about prompts, tokens, and temperature and understand them in detail.

6. Tokenization

In this lesson, we will discuss about tokens, which are the "currency" on the OpenAI models that process text by breaking them into smaller unites called tokens. We will understand how tokenization works in ChatGPT.

7. Temperature Control

Here, we will understand about a parameter of ChatGPT, temperature, which is a value anywhere from 0 to 1. The higher the value, more is the risk taken by the model and a lower value means the answer is a confident one.

1. OpenAI API Account Sign-Up

This video demonstrates how to create an OpenAI API account and provide prompts to models like Davinci and ADA, which are prompt-response-based application models. We will create an API key.

2. Azure Account Creation

The next step in the prework process is that we will set up an Azure account, which we will need to use Azure Functions and which will allow us to execute code without a server. We will need to connect to a server to get the prompt-responses, which will now be handled by Azure.

3. Azure Function Resource Creation

After learning how to create the Azure account and creating one, we will now create our first Azure function resource. We will click on the create tab and provide details required to create the Azure Function.

4. Python Installation

We will now download and install Python and set it up on our local system. We will download and install the Python 3.9.13 version and add the path for the software. We will check the installation using the CMD command.

5. Visual Studio Code Installation

The next thing we need to set up in our system is the Visual Studio Code, an IDE available by Microsoft, to deploy functions on Azure. We will create a blank project on a folder of choice.

6. Visual Studio Code Configuration

In this video, we will continue with the previous demonstration where we will need to create a few more configurations to our Visual Studio Code to get it up and running for creating our Azure Functions. We will install the recommended extensions for Python.

7. Terminal Use: Running Python and Pip

Here, we will verify that our Python commands are acceptable on the terminal. We will open the terminal and launch Python. We will define the path of the environment variables to help recognize the codes and commands.

8. Microsoft 365 E5 Account Creation

In this video, you will learn to create a Microsoft 365 E5 account that we will need. We will use the free trial option. This is needed to set up Azure. If not, you can alternatively use Tempmail to create an account.

1. Python Access to OpenAI API

The next stage is to create a serverless ChatGPT API, which enables us to ping OpenAI API. We will build Azure functions that will ping the API, ask questions, and show the responses on the API. We will link out Python to the OpenAI API to do this.

2. ChatGPT Response Management

Now that we have a way to sue a Python script on the API, we ran a script, which was accessed by the API, and returned a response, which is a JSON file. We will create a variable "Output" that will present the responses of our script.

3. Basic Azure Function Creation

In this video, we will create an Azure function. We will create a new project folder and create our first Azure function. We will create the function on our local workspace and then push the function over to Azure.

4. Azure Function Testing and Deployment

After creating our first Azure function, we will test it by deploying it so we get some practice and then look at how it works so we know how an Azure function works when it actually becomes truly serverless.

5. Postman for Serverless Function Testing

Now, we will create our Postman account and sign-in to our Postman account to deploy our Azure function. We will create a new collection called APIs and create a new request with an API. We will test the API using Postman.

6. OpenAI ChatGPT and Azure Functions Integration

After learning to deploy an Azure function using Postman and how to deploy an OpenAI API, we will now work on integrating the two together. We will call and deploy a function and learn to format the response for better readability.

7. Local OpenAI ChatGPT Testing

In this video, we will test our Azure function over a local workspace for functionality. We will define the parameters for the response and then test the function by deploying it. You will learn to create a wrapper API and integrate it into other functions.

8. Deploying OpenAI ChatGPT to Azure Functions

After successfully testing and deploying our function in the local workspace, we will now test and deploy our function in the OpenAI ChatGPT to create responses for the Azure functions that we have created.

1. Power Automate Flow Creation

After learning to create our serverless ChatGPT API, we can plug this into different software. We will look at integrating the API into Power Automate to begin with, which creates a flow for repetitive business tasks.

2. HTTP Request: OpenAI ChatGPT Integration

In this video, you will learn how to create a flow on the Microsoft Power Automate software. You will learn about the different parameters, passing URLs. We will create a JSON text request for the body of the flow.

3. Creating Teams Chatbot with OpenAI ChatGPT

After learning how to create a simple Flow on Power Automate, we will now use Power Automate and Teams to create a chatbot to integrate it with ChatGPT. We will ensure to use the same login for Power Automate and Teams.

4. Testing and Publishing Teams ChatGPT Chatbot

After saving the flow with the parameters defined in the previous lesson, we will now test the flow manually and we will note the response generated by the Power Automate flow on to the Teams message window.

2. OpenAI ChatGPT Integration in Power Apps

We will now create an app that sends a prompt to the Power Automate, and then sends the prompt to our API that we created and then to the Davinci model of OpenAI, which then sends the response from the Azure function to the Power Automate, back to Power App.

3. Power Apps Q&A Bot Implementation

We will now link our Power Automate flow to have the ChatGPT APIs into Power App. We will create the functionality when clicking the button to execute the flow and obtain the response from the Power Automate flow.

1. Office Add-In Environment Setup

The next common software that we will integrate OpenAI with is Outlook. We will use an Outlook account with Office products. We will begin with installing the tools and setting up the environment.

2. Office Outlook Project Creation

The next step involves creating our add-in project. We will use the Yo office command to select the support to Outlook. We will check to define the parameters that will modify the behavior of the app.

3. Outlook Add-In Launch

After building our first Outlook add-in project and going through some of the code behind it, we will start it and look at it on Outlook. We will be creating a task pane and we will click a button and a pane will come up from the right side.

4. Running ChatGPT API from Outlook

After the Outlook add-in is working, we will now create API calls from our Outlook add-in to the ChatGPT API that we created. We will make post-requests and edit the model, prompt, body, and temperature of the Azure code.

5. Resolving CORS and Security Errors

This video is for those who were not able to connect the Outlook with the Azure function. We will discuss the CORS error or any other security errors that might pop up during the testing and resolve the errors.

6. Email Body Transfer to ChatGPT

After getting the Azure functions and Outlook integrated, we will now create a useful application, which will allow users to get into their email and summarize the email for us. We will send the content of the mail and receive a response summary in the pane.

7. Pasting Response in Outlook Add-In

In this video, we will use the Outlook pane application we created and send the body of an email for which we require a summary of, and then generate a summary of the email and display it on the pane we created.

2. Bubble-ChatGPT API Connection

In this video, you will learn to place a text element, an input element, and a button element. The user feeds his input in the input element and clicks on the button, the button click sends the API across, and displays the response in the text element.

3. Dynamic Prompt Modification

In this lesson, we will understand how to create a dynamic prompt; the prompt should be equal to whatever input the user puts into the input A. We will click on the button and type in a prompt and await the response obtained and displayed in the Text element.

1. Airtable Context Setup

In this video, you will learn about creating the context for Airtable, which is a platform that creates databases and tables and tracks everything a business does and it has a built-in automation.

2. Script for ChatGPT API Calling

After creating the Airtable database setup, we will write script to generate emails. We will then integrate automation into the setup, add triggers, and create automation flows and generate prompts when there is a record update in the database.

3. Variable-Based ChatGPT Prompt Modification

After testing the script, we will edit the prompts section to correspond to the email we want to write based on the person's interest and interesting products. We will create a variable-based prompt and learn to modify it based on the responses.

4. ChatGPT Marketing Email Generator Testing

After completing the mail generation function, we will test the app. We will click on the generate mail checkbox and await the response generated by ChatGPT.

1. Front-End Setup

In this video, you will learn to create the frontend or the interface for the user to input data or destinations for the TravelPlan project. This will be the frontend for the TravelPlan project.

2. Second Page Creation

We will now create the second page for our TravelPlan project. We will create an input box for the user to add his travel plan and a box for displaying the itinerary generated.

3. ChatGPT Workflow Creation

Here, we will begin with creating the API called the Completion API as an action and the datatype to return as text. We will add the prompt, body, and URL parameters for the workflow.

4. Linking Pages and Testing

We will create a workflow where in the data is transferred to the second page of the TravelPlan project. We will save the data from page 1 into a database and then create a second session when clicking the page 2, where the data is brought up in a field.

5. ImageAI and OpenAI Photo Integration

In this video, you will learn to integrate ImageAI and OpenAI into the TravelPlan project. It will generate images and photos based on the location provided in the TravelPlan and make the itinerary page appear more interesting.

6. Text and Photo Testing for ChatGPT App

In this video, we will test our photo and images when providing a destination in the TravelPlan page. This will call the Completion API and ImageAI, which will produce not only the text itinerary but also interesting images of the location.

1. Outlook Environment Setup

We will create an app, "PolitelySayNo", which will reply to emails that you receive in your inbox. You will need to have Visual Studio up and running and also the Yo software available, and base it on the project where we linked Outlook and ChatGPT earlier.

2. Front-End Changes

After getting the basic setup up and running, we will begin with creating the frontend for our application. We will make a few changes to the task pane to display the name of the app, and other changes to how the frontend appears.

3. ChatGPT Back-End Function Creation

We will now create the backend for our application. We will create a button that, when clicked, reads the email and then uses the Completion API and generates a professional email, and sends the response back to the pane on the mail as a reply.

4. Body Modification for ChatGPT Backend

After testing the functions of the backend we created in the previous lesson, we will now learn to make modifications to the body that will alter the response for the mail generated in the pane.

5. Adding Reply Box

In this video, you will learn to create a reply box, which will read the mail received and generate a response or reply that can be used to email back the sender as a reply for the mail received from them.

6. Expanding Application with ChatGPT Mood Options

In this video, you will learn to make changes to the replies generated when clicking the Generate Reply button; instead of the standardized reply, we will be able to customize the reply with more information and modify the reply.

1. Cover Letter Screen Creation

In this video, you will learn to create cover letters for the JobApplicationCreator app, which will create a cover letter based on the post of the job applying for, which can be mailed to the employer.

2. ChatGPT API Linking and Workflow Creation

In this video, you will learn to create a workflow. We will create buttons and actions to the button and we will set up the API, which will use the Completion API that we created in the earlier projects.

3. ChatGPT Testing with Different Descriptions

In this video, we will test out our cover letter creator using different job descriptions to check the accuracy and professionality of the cover letter created by the application.

1. Image API DALLE Use

We will work on using the Image API DALL-E and integrate it an app called SplashImages, which will create characters based on images in a photograph; we will use an HTTP trigger and an Azure Function.

2. Front-End Creation

In this video, after creating the custom front-end API, we will check whether the API works by creating a request in Postman and updating the parameters of the prompts, body, and URL.

3. ChatGPT API and Power Automate Integration

We will now create the Power Automate script and create a new flow. We will integrate it with ChatGPT. We will create an HTTP request and the URL from Postman and Azure Function created as ImageAI.

4. Stock Photo Generator Testing

In this video, we will finally test the stock photo generator. We will run the application and verify the characters generated based on the text or image inputs that we provided and check the output based on the prompt request.